Source: ruby-rspeak Version: 1.0.4-1 Severity: serious Justification: FTBFS on amd64 Tags: sid Usertags: ruby2.7-transition
Hi there, the newly accepted ruby-espeak fails to test successfully with Ruby2.7. The reason seems that CSV.parse() has a behavior change. It does not parse the file test/fixtures/voices.txt anymore as it did. It seems using col_sep: ' ' is now interpreted literally. So fields separated by more then one space are not parsed as columns in a row. To demonstrate consider this line: > Pty Language Age/Gender VoiceName File Other Langs > 5 af M afrikaans af With ruby2.5 it is parsed like this: > #<CSV::Row "Pty":"5" "Language":"af" "Age/Gender":"M" "VoiceName":"afrikaans" > "File":"af" "Other":nil "Langs":nil> and with ruby2.7 like this: > #<CSV::Row "Pty":nil "Language":"5" "Age/Gender":nil "VoiceName":"af" nil:nil > nil:nil nil:nil nil:nil nil:nil nil:nil "File":nil nil:nil nil:nil nil:nil > nil:nil nil:nil nil:"M" nil:nil "Other":"afrikaans" "Langs":nil nil:nil > nil:nil nil:nil nil:nil nil:nil nil:nil nil:nil nil:"af" nil:nil nil:nil > nil:nil nil:nil nil:nil nil:nil nil:nil nil:nil nil:nil nil:nil> This seems intentional: https://github.com/ruby/csv/issues/67 https://github.com/ruby/csv/commit/7798df60fed87251b26c1202eb251a7894b55469#diff-fd263cdff2717a557bddf1592762dba3R16 The file format is determined by the output of `espeak --voices` and cannot be changed. Does anybody know, how to easily fix this, or is anybody up to add some magic to lib/espeak/voice.rb to deal with this? Regards, Daniel
signature.asc
Description: This is a digitally signed message part
